基于AI大模型的新能源汽车智能座舱多模态交互技术研究综述  

A Survey on Intelligent Cockpit Multimodal Interaction Technology for New Energy Vehicles Based on AI Large Model

作　　者：邓建明 龚循飞 于勤 罗锋 廖程亮 Deng Jianming;Gong Xunfei;Yu Qin;Luo Feng;Liao Chengliang(Jiangxi Isuzu Motors Co.,Ltd.,Nanchang 330108)

机构地区：[1]江西五十铃汽车有限公司,南昌330108

出　　处：《汽车文摘》2025年第1期8-13,共6页Automotive Digest

摘　　要：智能座舱是汽车智能化的重要组成部分,多模态交互是智能座舱的核心功能。为了研究AI大模型赋能智能座舱实现多模态交互的技术原理,利用AI大模型的学习和泛化能力,分析了多模态交互技术框架和关键技术,评估国内外科技公司和车企在多模态交互领域的应用案例,如百度、华为、腾讯和科大讯飞的大语言模型,并对其效果进行对比。对比结果表明,AI大模型在多模态交互中的应用显著提高了任务处理效率与准确性,增强了智能座舱的人机交互体验。最后探讨了AI大模型在智能座舱应用中面临的挑战与前景,为AI技术在智能座舱领域的深入发展和应用提供参考。The intelligent cockpit is a crucial component of automotive intelligence,where multimodal interaction serves as a core function.To explore how AI large models enable smart cockpits to achieve multimodal interaction,this paper leverages the learning and generalization capabilities of AI large models to analyze the technical framework and key technologies of multimodal interaction.It evaluates application cases from leading technology companies and automakers,including Baidu,Huawei,Tencent,and iFlytek,while also undertaking a comparative analysis of their performance.The findings reveal that the incorporation of AI large models into multimodal interaction has markedly enhanced task processing efficiency and precision,thereby enriching the human-machine interaction within intelligent cockpits.In conclusion,this paper explores the challenges and future directions for the application of AI large models in intelligent cockpits,offering valuable insights for the ongoing advancement and integration of AI technology in this domain.

关 键 词：新能源汽车 智能座舱 多模态交互 AI大模型 技术框架 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]